JupyterHub and BinderHub Team Meeting - July#
Date: Thursday 16th July, 2020
Time: 5pm UTC
GitHub issue: jupyterhub/team-compass#308
Calendar for future meetings: https://jupyterhub-team-compass.readthedocs.io/en/latest/meetings.html
Welcome to the Team Meeting#
If you are joining the team video meeting, sign in below so we know who was here. Roll call:
name / institution / GitHub handle
Min / Simula / @minrk
Sarah / Turing / @sgibson91
Hamel / GitHub / @hamelsmu
Chris / Berkeley / @choldgraf
Matthew / Illinois & PyHEP / @matthewfeickert
Tim / Binder/ @betatim
Steve / AWS / @blink1073
Darian / Two Sigma / @afshin
Georgiana / CIR / @GeorgianaElena
Richard / Aalto / @rkdarst
Erik / Sundell Open Source / @consideRatio
Rollin / NERSC / @rcthomas
Michael / Code for Philly / @machow
Simon Li / OME, University of Dundee / @manics
Zach Sailer / Jupyter Cal Poly / @Zsailer
60 second updates on things you have been up to, questions you have, or developments you think people should know about. Please add yourself, and if you do not have an update to share, you can pass.
Hamel: repo2docker GitHub Action, allows you to dockerize your projects and cache them for Binder from GitHub!
We had a bunch of Binder usage over the last two weeks - first from SciPy and then from PyHEP 2020. Both have gone pretty smoothly, but brought up some questions about our process around catering to specific events. Can find a PR and several links to issues here: jupyterhub/mybinder.org-deploy#1524
Chris is going on paternity leave soon-ish. The babypocalypse is nigh! Expect him to be totally unreliable roughly mid-August through September (Chris promises to keep approving Numfocus payments for our Contributor in Residence :-) )
Chris has been working to launch a new non-profit with some others in the community. It’s called 2i2c (https://2i2c.org/) and is just getting started. It will try to find a way to sustainably offer jupyterhub environments to research and education organizations, and provide support for our repositories / communities as well.
Steve: Jupyter Server migration
Erik: long battle with a reliability issue with Z2JH’s automatic TLS certificate acquisition
Reports and celebrations#
This is a place to make announcements (without a need for discussion). This is also a great place to give shout-outs to contributors! We’ll read through these at the beginning of the meeting.
Sarah: mybinder.org federation is now upgraded in parallel by GitHub Actions! Thanks to Min for helping with the deploy and debug.
Chris reverses this “thank you” back to Sarah for doing all of this great work!
Chris: many thanks to Carol and Tim for helping debug what was going on with the xarray tutorial during SciPy!
Let’s collect all potential agenda items here before the start of the meeting. We will then attempt to create a coherent agenda that fits in the 60m meeting slot. If there are similar items try and group them together.
Hamel Husain (10 min): Proposal to include the repo2docker Action in the Jupyter org: issue
Tim can move the repo in terms of permissions
check if there are paperwork issues to do first
contributing guide, COC, license, etc
people liked the idea of having it migrated to the JupyterHub org
conversation will continue in the issue
used to take a HTML page and hook it up to a kernel running on a JupyterHub like mybinder.org
the tech stack for thebelab is more JS than Python/Kubernetes, does the Hub team have the skills to keep maintaining it?
not a huge amount of work to maintain
biggest source of “work” is keeping up with Jupyter Lab updates
some outstanding PRs and Issues that would be nice to adopt
people liked the idea of having it migrated to the JupyterHub org but basically split down the middle on executablebooks vs jupyterhub
discussion continues in the issue linked at the top
Matthew (10min): Binder does PyHEP - retrospective
we bumped the limits for two or three repositories each day of this week
how did it go? What went well, not so well, random?
Feedback: it was hard to convey to the authors of talks what binder is and why they need to prepare links more than 5min before the talk (Matthew: I’ll note that this wasn’t really any problem with Binder or the way that the docs are - they’re great!. This was more just getting people to read, which I’m unqualified to offer solutions for.)
Q: What did the audience think/say about having a binder?
People asked interesting questions based on being able to run the code
takes a moment for people to warm up to the idea, kept telling them to give it a go (Matthew: People seem excited and happy to use it once they’ve done so. More just that when you have 1000 people that are connecting for different portions you need to repeat everything again and again for the first 3 days.)
Q: did people use the video links provided in a binder?
probably not so much as the conference already provided zoom, slack, etc
Matthew: Standing offer from the PyHEP organizers for a Binder keynote at future PyHEPs. :)
Tim, 5min: Microsoft Quantum Computing people make good use of mybinder.org
the quantumkata repo shows up on my radar once in a while
should we use the request to bump resources as a way to ask for cloud credits/sponsorship on Azure?
we have several contacts and we should start having discussions with all of them
there might be talks from core projects already w/o proposals, but check with Jason Grout
Rollin (5 min): Does anyone run Prometheus+Grafana as part of their hub deployment with Grafana as a JupyterHub service (admin access only) and are there any potential gotchas (I have a test deployment…)
rkdarst: I do as part of my k8s hub, but I haven’t thought about it much and am interested in practices. I have an extra service that exports more data, and runs a periodic test job link - it’s very hackish.
Erik: (k8s) Yepp. Admin access only is not integrated with JupyterHub authentication or similar atm though, but simply as a pre-created admin user. The configuration of this Helm chart that depends on the JupyterHub helm chart, Grafana Helm chart, and the Prometheus Helm chart, and follow a structure as required by the tool called
AWESOME thanks! Rollin will write something up and post it on discourse
At what point should things be offered for adoption by the jupyterhub organization?
Balance: increased visibility vs increased team maintenance burden
Are the people of JupyterHub interested in maintaining it?
Considerations for issue template:
are the people who created the project coming along to continue maintaining it?
Does it have good tests?
TODO: create an issue template to help make this process more repeatable and easier
What about de-adopting projects if it turns out to be a bad idea… / it goes stale?
maybe use badges or a note in the README to signal that a repo is in need of new maintainers
check out https://repostatus.org